Sie sind auf Seite 1von 6

Recent Advances in Finite Differences and Applied & Computational Mathematics

Thermal analysis of the rotary kiln through FEA


SUSANA ARAD
Computer, Informatics and Electrical Engineering Department
University of Petrosani
332006 Petrosani, 20 Universitatii str.
ROMANIA
susanaarad@yahoo.com, http://www.upet.ro
Abstract: - This paper aims at analyzing numerical heat transfer in kiln clinker production in cement factory
Deva, important processing equipment in the cement production. The importance of knowing this process aims
to rationalize the consumption of raw materials and energy but also to get high standard finished products. We
used QuickField software package Version 5.10.1 for analysis of thermal problems in a rotary kiln to obtain an
optimal and efficient sinterization process. In this study, our objectives are to predict the temperatures in the
inner surface/refractory and inner surface of bed before any variations in process parameters, to decrease the
errors arisen from the operators, to increase the efficiency and finally decrease the process cost also the
variation of inside temperature with flame temperature.
Key-Words: - FEA, modelling, heat transfer, rotary kiln, efficient operation, clinker.
impact on the environment which is why a lot of
investment projects headed in this direction.
Process automation and information technology
are industrial areas which offer the biggest and most
satisfying challenges in terms of combining
traditional engineering skills with technological
innovation [4].
The demands on cement industry in relation to
productivity, quality and price, means an ever
increasing need to improve the quality products, to
productivity increase improvement of products
quality, modernization of the technological flow and
environmental quality.

1 Introduction
Today, Portland cement is the most widely used
building material in the world with about 1.56
billion tones (1.72 billion tons) produced each year
(Cement Association of Canada, 2001). Global
economic situation sets more and more challenges
to cement manufacturers both in our country and
globally. Although the cement market in Romania
has reached a high level with annual an increase up
to 30%, the end of 2012 was faced with stagnation
or at most an increase of 1-2% and in 2013 most
likely cement market will record a standstill,
because of demand, but also due to rising prices. It
must be noted that the need to maintain producers
on the market is due to more and more discipline
regarding costs and optimizing them according tp
technological efficiency and rationalization of raw
materials and energy (Carpatcement Holding, 2013)
The main cement producers on the Romanian
market are: Carpatcement (Bicaz factories, Fieni
and Deva), Lafarge (Medgidia Hoghiz and Targu
Jiu) and Holcim (Turda and Alesd). Carpatcement
Holding hope for a rearrangement of the
macroeconomic situation of the country starting this
year and also for a market that depends on several
conditions so that construction work can be started
earlier than in 2012 (Carpatcement Holding, 2013).
The cement industry is actively engaged and
committed
to
sustainable
developmenta
philosophy that focuses on meeting our construction
needs today without depleting future resources.
Cement industry production process has a big

ISBN: 978-1-61804-184-5

2 State of Art
Portland cement clinker is produced from a mixture
of raw materials containing calcium, silicon,
aluminum, and iron as the main elements. The
mixture is heated in kilns that are long rotating steel
cylinders on an incline. The feed of a typical cement
kiln consists of limestone and other, mainly oxide
and silicate based material, which typically contain
calcium, magnesium and iron. The process goes
through the raw material grinding, blending,
precalcining, calcining (clinker formation), cooling
and grinding of final product. In the preheating
system, the raw material goes through preliminary
steps, where the calcination of CaCO 3 and
formation of calcium oxide takes place.
The material is fed from the top of the cyclone
and move downward in counter stream with gas

182

Recent Advances in Finite Differences and Applied & Computational Mathematics

monitoring using integrated systems there are many


objectives which can do the efficiency growth.
Using the PLCs and expert systems has been
brought forward to the board of the company,
managing to deploy an expert system. Thus the
technological process is carried out from the
command room using the visualisation system ECSNtech and the PLCs Allen Bradley (kiln and cooler)
and Siemens [12]. Through this system the attendant
can overlook the performance of the technological
process using the computer network. [1]
Correct measurement of the temperature on a
kiln shell is essential for efficient operation of the
kiln. ECS/CemScanner represents the state-of-theart in kiln shell infrared scanning. The thermal
detailed image and real-time monitoring of the ring
formations in the kiln are set up. If obstacles prevent
the scanner from viewing parts of the kiln shell, up
to 8 infrared pyrometers can be installed to measure
temperatures in these specific areas. [12] To obtain
an optimal and efficient sinterization process both in
terms of economy of materials and to obtain quality
products, besides all the monitoring and control of
rotary kiln temperatures we performed a thermal
analysis using software QuickField.

resulting from the combustion process that flows


through cyclone separator from lower levels to
higher ones. Flow chart of clinker production is
shown in Fig. 1, [12]. In this way the gas are cooling
as raw material preheating and start the process of
calcinations. Raw material in suspension separates
the gas in each gear shift and gas reunited in the
descent step. This process are repeating (separation mixing) every step of unloading mix material
exchanger to ensure good furnace heat transfer.
The calcination (transforming of raw material
into clinker, the intermediate product used to
produce cement) occurs in rotary kiln according to
the four zones: calcination area, transition zone,
clinkering zone and cooling zone. The kiln
temperature needed for this process is about 1450 C
and flame temperature of approximately 2000 C. [5]
The rotary kiln is divided into successive control
volumes or cells [6]. The volume elements of the
bed and the gas in the cells are described as open
thermochemical systems, which transforms heat and
mass with each other.

3. Background
The finite element method FEM is used to solve
complex engineering problem [2]. Engineers who
need to model electromagnetic or thermal fields
frequently turn to finite-element analysis (FEA), a
numerical technique for solving field problems of
all kinds [9]. The first step in finite-element analysis
is to divide the analyzed configuration into small
homogeneous elements. The model contains
information about the device geometry, material
constants, loads and boundary constraints [3]. In
each finite element, a linear variation of the field
quantity is assumed. The corners of the elements are
called nodes. The goal is to determine the field
quantities at the nodes. The Finite-Element Analysis
technique solves the unknown field quantities by
minimizing energy functional. The energy
functional is an expression describing all the energy
associated with the configuration being analyzed.
QickField is a finite element analysis software
package, a slick FEA for solving two or three
dimensional field problems. From studies and
evaluations conducted by a team of developers it is
estimated that QuickField is easy and suitable for
this problem. [11]
Compared to other FEA programs, QuickField
has two main facilities: exceptional user-friendliness
and blinding speed. The program has a typical

Fig. 1. Clinker production flow chart [12]


The clinker leaving the kiln enters the cooler
rack which has two functions: to recover a larger
amount of energy from the hot clinker to be used in
the process and to reduce the clinker temperature to
the right level for its later semi fabricated use for
cement production.
Although there is a quite high level of
automation and the control of installations for each
stage of cement production from Deva Factory and

ISBN: 978-1-61804-184-5

183

Recent Advances in Finite Differences and Applied & Computational Mathematics

Windows interface, drop-down menus, and


extensive help facility, and is equipped with many
sample programs [11]. While the user has to
understand something about the physics of the
problem to successfully carry out a simulation, the
mechanics themselves are straightforward. Even on
a slow PC, users can construct and run models with
tens of thousands of nodes, which make it practical
to solve very complex models on modest platforms.
The program itself is small, a few megabytes with
various required files.
Heat-transfer equation for linear problems in
rotary kiln is represented by the differential
mathematical model of the thermal conduction [11]:

Fig. 2 Heat transfer mechanisms from the kiln

div(gradT ) + q c

T
=0
t
,
T
T T
= q c
+ y
x
t
x x y y

Convection (from gas to bed and inner wall and


outer wall to surroundings)
Conduction (between bed and inner wall and
inner and outer wall)
Radiation (main heat transfer mechanism).
A typical operational case of the kiln is modeled
to show the process inside the kiln and the
capability of the process modeling. The kiln is
fuelled by natural gas, liquid fuel and coal and solid
alternative fuel or fuel combination. The
combustion air is provided, which works out to
about 15% excess air. 80% of the total combustion
air is secondary air, and the rest goes into the kiln
through the gas burner.
The model developed was for a rotary kiln of 5.8
m in diameter on a vertical cross section with 0.58
cm diameter. As mentioned previously the model
takes into account convection, radiation and
conduction from combustion and from the hot gas.
A combustion profile is incorporated into the model
to simulate the non instantaneous combustion
process. The parameters of rotary kiln from Deva
Factory are: diameter of 5.8 m, length of 97 m,
production capacities Q = 3000 t/ day, the main
driving power, P = 500 kW with rotation speed, n =
750 rot/ min, second driving P = 500 kW with n =
750 rot/ min and an angle from horizontal position
of 3 %.

(1)

(2)
where: T scalar temperature T; t - time; x (y,r,z) components of heat conductivity tensor; (T) - heat
conductivity as a function of temperature
approximated by cubic spline (anisotropy is not
supported in nonlinear case); q (T) - volume power
of heat sources, in linear case - constant, in
nonlinear case - function of temperature
approximated by cubic spline; c (T) - specific heat,
in nonlinear case - function of temperature
approximated by cubic spline and - density of the
substance.

4. Modelling of heat transfer with FEA


The clinker production process involves heat and
mass transfer between the kiln, fuel, primary and
secondary air, drying of raw material mix call flour,
and calcining of CaCO 3 [6]. It is important to have a
comprehensive understanding of these processes in
order to: diagnose operational problems, improve
energy consumption, increase production, reduce
emissions, increase refractory life, improve the
product quality and optimize kiln operations.
The given computer model of thermal processes,
allows to calculate temperatures pattern both at a
surface and inside a body [10]. The model allows to
perform calculation heat transfer of a multy-layer
ring structure of a body with a granular mixture clinker, roasted inside it (Fig. 2). [7] The kiln body
is coated with refractory brick and is analysing like
a multy-layer body: shell, chamotte, air inside and
clinker. There are three different heat transfer
mechanisms in the rotary kiln, Fig. 2 [7].

ISBN: 978-1-61804-184-5

4.1. Model formulation


The first stage of Finite Element Analysis is
problem description that includes: edit problem
description options, as well as material properties,
field sources and boundary conditions. Our problem
is a steady-state heat transfer, linear with
axisymmetric model class. In our linear problem all
the parameters are constant within each block of the
model.

184

Recent Advances in Finite Differences and Applied & Computational Mathematics

source is concentrated and defined by a block unlike


the second option with four blocks in which the heat
source is defined by a labeled vertex with boundary
condition and source specified through temperature
value T = 2000K.
Edge label with boundary condition specified for
two scenarious are shown in Table 2.

The model is described through a multi layered


with different properties model, that is the outside
layer is formed of 2 layers and the inside of the kiln
contains the gas, clinker and heat source. The
geometric model is composed of blocks with
different properties, in this case, it was assigned 4
blocks with material properties specified (air,
chamotte, clinker, steel shell), Table 1. In each
block the material properties was specified (thermal
conductivity and volume power of Heat souces Q).
The model geometry was described by create
vertex label with source specified (burner flame Q)
and edge label with boundary condition specified
that form boundaries of all subregions having
different physical properties. We created two
geometrical models.
Table 1. Material properties in each blocks
Block Label
Thermal conductivity
1 (W/K.m)
2 (W/K.m)
air
11
11
chamotte
0.165
0.165
clinker
1.2
0.1225
steel shell
22
22

Fig. 3 Geometrical model and the mesh


Table 2. Boundary conditions applied
Edge Labels
Temperature
Temperature
T 1 (K)
T 2 (K)
Outer
293
293
Inner
0
0
Bed surface
1450
1000

Applied mesh is shown in Figure 3 and the field


values at the nodes have been calculated.
Finite Element analysis is based on discretization
of the area into the number of simple shapes, in case
of QuickField triangles. Accuracy of the results
highly depends upon the finite element mesh density
and quality. Moreover, in complicated cases
problem solution on the non-adequate meshes may
not converge. Thats why the mesh generation is
one of the most important operations in all practical
FEA applications. QuickField Model Editor is
equipped with very efficient mesh generators. But
sometimes, then the geometry was really complex,
smooth mesh generation required many manual
adjustments. Improved mesh building algorithms of
this version 5.10 assure automatic generation of the
smooth meshes even in the most complicated cases,
which allows good results of the model.
We have built the mesh in all blocks
participating in field calculation. Flame positioning
for example can be predicted in multi-dimensional
models rather than used as an input. In one of the
models the flame position is concentrated in the
center of the section and in a point of the upper half
of the section in other model, Fig. 3. Each model
was individually tested and where appropriate, was
compared to experimental results.
The first option has been created with a
digitization network of 5 blocks in which the heat

ISBN: 978-1-61804-184-5

Because thermal conductivity function of


temperature in nonliner case, we have carried out
two scenarios considering the conductivity
variation, = (T) of clinker and results are shown
in Fig. 4 for 1 and in Fig. 5 for 2.

Fig 4 Temperature distribution for 1


The energy to raise the temperature and drive
endothermic reactions is from the combustion of a
range of fuels such as natural gas, coal and more
and more alternative fuels.

185

Recent Advances in Finite Differences and Applied & Computational Mathematics

kiln. The temperature profile at inside the kiln


follows a parabolic function [7].

Fig. 5. Temperature distribution for 2


Fig. 7. Temperature distribution for 2, T2=1000

4.2 Temperature distribution inside


The feasibility of simulating heat transfer 2D using
a common, finite element modeling package
QuickField was tested. The aim was to recreate well
known, but difficult to simulate, interactions
between thermal transfers in multy layer structure.
For the test application, QuickField was chosen
for the following reasons: the Heat Transfer module
is used to analyze the temperature distribution in
static and transient heat transfer processes; nonlinear
or anisotropic properties; boundary temperature and
heat fluxes and boundary conditions with
convective/radiative terms.
Temperature distribution T, in vertical section
and vectors of Heat flux F are presented in Fig.5
when it was chosen a temperature inside about
1450C for surface of bed.
The contour definite for calculus of quantities
value is shown in Fig 6 and also isotherms.

4.3 Results
To analyze the results obtained on the contour
defined in Fig. 6 and Fig. 7 at temperatures T1 and
T2 we have performed parametric analyzer with
LabelMover. We have specified the values which
we want to optimize and start optimization process
for some physical quantities.
We have calculated with Harmonics Browse the
plot for temperature values and amplitudes for the
two boundary conditions, T1 and T2. Harmonic and
linear approximation of the temperature variation
(K) on contour inside the kiln (m) that corresponds
at the whole period, are shown in Fig. 8. The plot of
temperatures approximation for the value of 1
without conditions imposed for the temperature of
clinker bed surface, only temperature source, like
T= 2000K, has been raised.
1732.7

T e m pe ra tu re

1688.5

1644.4

L in e a r a ppro x im a tio n
H a rm o n ic a ppro x im a tio n

Fig. 8.Plot for temperature approximation


This add-in allows us to calculate and view
harmonics (that is, phases and amplitudes for
Fourier series) for any displayed quantity related to
the currently selected contour.
The graph of temperature variation, obtained
from model described has been presented in Fig. 9.
We stopped at the results that are comparable with
the results achieved from measurements. The
following results are retrieved: temperature, heat
flow, thermal gradients, and total heat loss on any

Fig. 6. Temperature distribution for 2, T1 =1450


The same contour was chosen for temperature of T
= 1000K and quantities value has been collated, Fig
7. We have applied two values on the boundary
conditions for the temperature inside the kiln on the
clinker surface T1 and T2. The combustion to the
fuel is assumed to occur at a point two meters into
the kiln and is complete at a point 32 m into the

ISBN: 978-1-61804-184-5

186

Recent Advances in Finite Differences and Applied & Computational Mathematics

given part and other whole inner surface of


quantities for clinker bed.

source can be changed to accommodate longer less


intense flames. To maximize refractory life, it is
essential to avoid flame impingement on the
refractory lining. Therefore it is recommended to
positioning burner system in order to avoid possible
refractory failure in that region. Engineers can use
the software to diagnose equipment problems.

References:
[1] Arad, S, Arad, V., Heat transfer in rotary kiln
from Deva cement factory. Papers SGEM2012/
Proceedings, ISSN 1314-2704, Vol. 2, 2012, pp
373 380;
[2] Bargallo, R., Finite Elements for Electrical
Engineering, EUETIB-UPC, 2006;
[3] Kenneth R. Foster, Quick and easy field
analysis, IEEE Spectrum, Vol. 36, Number 12;
[4] Marcu, M., Uu, I., Popescu, F., Pana, L.
Simulation Software for Static Switch
Controllers. Annals of University of Petrosani,
Electrical Engineering, Vol 9, pp 153-159,
Petrosani, 2007.
[5] Penttil, K. A Simulation Model of TiO2calcination Kiln. M.Sc. thesis, Helsinki
University of Technology, Faculty of Process
Engineering and Materials Science, 1996.
[6] Tscheng, S. and Watkinson, A. Convective Heat
Transfer in a Rotary Kiln. Can. J. Chem. Eng.,
57, pp. 433443, 1979
[7] Palmer, G., Howes, T. Heat transfer in rotary
kilns, Cement Industry Federation Technical
Conference 1998,
[8] Petrilean, D.C., Mathematical Model for the
Determination of the Non-Stationary Coefficient
of Heat Transfer in Mine Works, RECENT
ADVANCES IN MATHEMATICS (ASM '13),
Published by WSEAS Press, pp 124-130, 2013
[9] Powell, S., Finite element modelling of
magnetostatics for magnetron sputter sources.
"CAD for Electromagnetic Devices", seminar, 2
November 2006, Rolls Royce plc, UK
[10] Zerafat Angiz, F., Amanifard, N., Haghi, A. K
A Numerical Study on Thermal Drying of Moist
Porous Solid, Proceedings of the 5th
IASME/WSEAS INT. CONFERENCE ON
HEAT
TRANSFER,
THERMAL
ENGINEERING AND ENVIRONMENT,
Athens, Greece, 2007
[11] QuickField 5.10 User's Guide Tera Analysis
Ltd.
[12] CARPATCEMENT
HOLDINGDEVA
Factory, Dispatcher Report on ECSNTech software, 2012

Fig. 9. Temperature variation on bed surface


Data achieved from a full-scale cement kiln at
Deva factory are comparable to results achieved
with the software used. The best approximation is
achieved on the clinker surface when the
temperature of the heat source is taken from the
pyrometer inside the rotary kiln like in Fig. 10.

Fig. 10 Data from pyrometer measurements


inside the kiln

5. Conclusions
The validation of any model is extremely
importance for the effective use of the model as a
design tool. Validation to heat transfer has been
done using full scale kilns at Deva cement factory.
In further development of the model the aim is to
incorporate not only thermal profiles for the rotary
section of dry process kilns but the effect of chain
systems in long wet kilns including dust entrainment
through the chains.
We see significant temperature increase close to
flame of burner (e.g. T = 1810K). However, the heat

ISBN: 978-1-61804-184-5

187